TELSAM Fusion Protein Design

Protein Design of a short 1TEL-target protein fusion if target protein has an N-terminal a-helix 

1. First, a computational model of a short 7-subunit TELSAM helix is made using PyMOL . PDB ID 2QAR works well for this.
   
2. Second, the N-terminal-most stably positioned amino acid (NSPAA) of the target protein is identified. Usually this is the first hydrophobic amino acid on the target protein N-terminus that is packed against or amidst other hydrophobic amino acids. Target protein models can readily be found using the UniProt database.  
   
3. N-terminus amino acids upstream of NSPAA are deleted.  
   
4. Third, if the target protein has an α-helix at its N-terminus of if the NSPAA is part of an α-helix, that α-helix is superimposed (using the ‘pair_fit’ function in PyMOL) onto the C-terminal α-helix of N-terminal most subunit of the 7-subunit TELSAM helix model. 
   
5. Increasingly closer superpositions are made until the target protein exhibits steric clashes with the host TELSAM polymer. Some clashes can be mitigated by sampling alternative conformations of clashing amino acid side chains.  
   
6. Once the closest superposition has been identified, unneeded amino acids at the TELSAM C-terminus are deleted up to but not into the essential …DELYELLQHIL motif. A semi-rigid connection can be achieved by substituting one of the helical amino acids with proline, such as …DELYELLQHIL-P-target_protein. 
   
7. The target protein and 7-subunit TELSAM helix model are combined into one chain and exported for subsequent modeling using Coot. 
   
8. In Coot, side chain and backbone conformational sampling are used to ensure that the linker has physiological bond lengths and angles and to resolve any clashes between the target protein and the TELSAM polymer model.  
   
9. If needed, FolditStandalone can additionally be used to enforce physiological bond lengths and angles.  
   
10. It is important to ensure that the amino acids between the TELSAM …DELYELLQHIL motif and the target protein NSPAA are as bulky as possible.  
    
11. If additional amino acids remain in the linker downstream of the …DELYELLQHIL motif, they can be mutated if needed to resolve steric clashes with the target protein.  
    
12. Model multiple copies of the new fusion protein against a TELSAM polymer from PDB ID: 2QAR to make sure that the new linker does not clash with the polymer or other copies of the displayed proteins. 
    
13. The resulting amino acid sequence is reverse-translated and a DNA construct is designed using Geneious. 

Protein Design of a short 1TEL-target protein fusion if target protein has an N-terminal b-sheet or loop 

1. If the NSPAA is instead part of a β-sheet or loop, it is superimposed via its backbone atoms onto the lysine that immediately follows the …DELYELLQHIL motif in PDB 2QAR. 
   
2. Unneeded amino acids in both the target protein and the 7-subunit TELSAM helix model are deleted and the target protein and 7-subunit TELSAM helix model are combined into one chain and exported for subsequent modeling using Coot.  
   
3. In Coot, side chain and backbone conformational sampling are used to ensure that the linker has physiological bond lengths and angles and to resolve any clashes between the target protein and the TELSAM polymer model.  
   
4. If needed, FolditStandalone can additionally be used to enforce physiological bond lengths and angles.  
   
5. It is important to ensure that the amino acids between the TELSAM …DELYELLQHIL motif and the target protein NSPAA are as bulky as possible.  
   
6. If additional amino acids remain in the linker downstream of the …DELYELLQHIL motif, they can be mutated if needed to resolve steric clashes with the target protein.  
   
7. Model multiple copies of the new fusion protein against a TELSAM polymer from PDB ID: 2QAR to make sure that the new linker does not clash with the polymer or other copies of the displayed proteins. These should at least include three copies of the new fusion protein, two of which are laterally adjacent and two are vertically adjacent. 
   
8. The resulting amino acid sequence is reverse-translated and a DNA construct is designed using Geneious. The sequence can be extracted and copied from PyMol using the following code while the desired protein region is selected: print(cmd.get_fastastr('sele'))
   
9. It is also important to ensure that the …DELYELLQHIL motif has a glutamate at the second position rather than the wildtype valine, as this is the critical pH-dependent polymerization trigger. 

Protein Design of a long 1TEL-target protein or target-target fusion 

1. Create your complex as multiple chains in PyMol and identify the target termini you wish to join. 
   
2. Determine which route between target termini is the shortest logical path for the linker to follow. 
   
3. In Coot, create your long linker by adding amino acids from each end along the path determined in PyMol until they meet. 
   
4. If you intend to design more than one linker length, save out the file at the first linker length and then continue adding or subracting residues on the same model to reach you desired alternate linker lengths, saving out each as a seperate file before creating the next. Do all of the linker lengths before moving to the next step, because adjusting the length afterwards tends to cause problems in Foldit, and you will save time having to return to the original model and rebuild from scratch for every linker. 
   
5. In Pymol, adjust the numbering of the construct and make it a single chain. It’s easiest to adjust the numbering first using: alter sele, resi=str(int(resi)+[some-number]) Afterward you can change the chain IDs: alter sele, chain=”A” If needed, you should also remove any segment identifiers: alter sele, segi=””
   
6. In Coot, regularize the connection point. This may not be necessary as long as the linker ends are close enough together, as Foldit can recognize and connect the two linker halves together automatically. 
   
7. In Foldit, minimize the entire structure, focusing on the connection point. Start with the repack tool (for sidechains) over the full structure, then select only the linker (holding shift allows you to select all residues between two points) and use the minimize backbone and sidechains tool (usually labeled as W, depending on your Foldit version). Keep the linker selected for the next step. 
   
8. In Foldit, optimize the identities of the amino acids in the linker (the tool labeled “Y”). Then use the “W” tool once more on the linker. 
   
9. It is important to ensure that the amino acids in the linker are as bulky as possible while not being too hydrophobic.  
   
10. Model multiple copies of the new fusion protein against a TELSAM polymer from PDB ID: 2QAR to make sure that the new linker does not clash with the polymer or other copies of the displayed proteins. If the vertically adjacent polymers are slightly too close, you may model using PDB ID: 7TDY as it has a slightly larger helical rise. 
    
11. The resulting amino acid sequence is reverse-translated and a DNA construct is designed using Geneious.  
    
12. It is also important to ensure that the …DELYELLQHIL motif has a glutamate at the second position rather than the wildtype valine, as this is the critical pH-dependent polymerization trigger.